Volatility squeeze

Thesis (rejected): Names with compressed realised volatility, tight Bollinger bands and a contracted 60-day range are "wound springs" that break out with elevated forward volatility.

Code: research/squeeze_validate.py Report: research/squeeze_report.json (n=17,774)

Why it failed: Tighter compression predicts less forward volatility, not more. The relationship is monotonic and the wrong sign — the super-tight bucket (realised vol < 10%) moves only ~2.4 % over 20 days vs the borderline bucket (vol 18-22 %) at ~5.8 %. The "wound spring releases" intuition is the opposite of what the data shows on a 20-day horizon. Volatility is autocorrelated; low vol predicts more low vol.

Verdict: Removed from the live watch-list 2026-05. Long-volatility plays (straddles, strangles) on these names would systematically bleed.